Optimize the Fault Diagnosis of Rolling Bearings using Stochastic Resonance and Extreme Learning Machines

This paper addresses the problem of low fault diagnosis accuracy in rolling bearings caused by weak fault signals that are susceptible to noise interference. To solve this issue, a fault diagnosis method for rolling bearings is proposed, which utilizes Stochastic Resonance (SR) and an Extreme Learning Machine (ELM) optimized by the Northern Goshawk Optimization (NGO) algorithm. Firstly, taking the negative value of the signal-to-noise ratio of the output signal as the fitness function, NGO is used to adaptively optimize the key parameters of SR, and the optimized SR is used to enhance the weak fault signal. Secondly, extract the fuzzy entropy of signals of different fault types as fault features; Finally, the fuzzy entropy features are input into the extreme learning machine for fault classification to complete fault diagnosis. Through experimental analysis, the proposed method can achieve a relatively high accuracy rate of fault diagnosis.